RE: Co-ordination protocol and BPEL

Ricky,

> >The general case is
> >         perform provisional work such that you can later:
> >
> >         make it not so, if so instructed
> >         make it finally so, if so instucted
> >
> >(slot in as many synonyms for provisional, make it not so, 
> make it so 
> >as you please)
> >
> >so, I agree, you need at least two "handlers". Whether these are 
> >defined as mainstream and exception, or both as event handlers is 
> >mostly a matter of how the language works, I think.
> 
> But in BA of WS-Tx, there is no "provisional" concept.  The 
> case is ...
> 
>          Do and commit.
>          If error
>                  Undo by compensate
> 
> I don't see where is the "co-ordination" which I'm assuming 
> some sort of 
> 2-phase message exchange.

A BA "complete/completed" exchange concerns a provisional state because
there is still the possibility of the compensate, until either
compensate or close is sent. (There is nothing that says the undo is
only triggered in case of error - like BTP cohesions, the client side
can just choose to cancel some participants.)

A consequence is that the service shouldn't count the "completed" things
as final until the close has removed the possibility of a reversal. (Of
course, in practice, that depends on what these things are, how many
statistically get reversed, how difficult it is to apply the
compensation, how urgent to get started on any consequential
processing).



Peter

Received on Friday, 23 May 2003 04:57:29 UTC